Nothing personal? A system-internal syntactic change in ...kroch/hist-pdf/maling-goett11.pdf · Sigurjónsdóttir & Maling (2001) and Maling & Sigurjónsdóttir (2002) Questionnaire

Nothing personal? ��A system-internal syntactic change in Icelandic

Joan Maling Brandeis University

Anthony Kroch University of Pennsylvania

Sigríður Sigurjónsdóttir University of Iceland

Comparative Germanic Syntax and the Challenge from Icelandic��

February 24, 2011

Iceland (103,000 km2; population c. 317,000)

Eyjafjallajökull

The Icelandic New Construction

Active: Hún bað mig að vaska upp. she-NOM asked me-ACC to wash up

‘She asked me to do the dishes’

Canonical Passive – verb governing ACC

Passive: Ég var beðinn að vaska upp. I-NOM was asked-masc.sg. to wash up

‘I was asked to do the dishes’

Theme/Patient marked ACC

Theme/Patient marked NOM

Það var beðið mig að vaska upp itEXPL was asked me-ACC to wash up literally: “it was asked me to do the dishes”

intended: “I was asked to do the dishes” or “they asked me to do the dishes”

This innovative construction has become known as either the “New Passive” (Kjartansson 1991; Barðdal & Molnár 2003; Eythórsson 2008; Jónsson 2009)

or the “New Construction”/ “New Impersonal” (Sigurjónsdóttir & Maling 2001, Maling &Sigurjónsdóttir 2002, Maling 2006, etc.)

The Innovative Construction

Það var beðið mig að vaska upp itEXPL was asked me-ACC to wash up literally: “it was asked me to do the dishes”

intended: ‘I was asked to do the dishes’ or ‘they asked me to do the dishes’

In a 1999-2000 nationwide survey (M&S 2002), ��93% of surveyed adults found this sentence

completely unacceptable.

73% of adolescents found it completely ��acceptable!

Age-related variation

Earliest known report with unambiguous object case:

1959, 8-year old girl in Akureyri

Það var bólusett okkur itEXPL was inoculated us

‘They inoculated us’ or ‘We were inoculated’

Another woman reported that the construction was common in Akranes in the late sixties. (cf. Sigurjónsdóttir & Maling 2001; M&S 2002)

Spread of the New Construction

Skoðað verður miða við innganginn. checked will.be tickets-ACC at the.entrance ‘Tickets will be inspected at the door’ (Maling 2006:200, ex. (7))

2004 Sign posted at Háskólabío movie theater

Spread of the New Construction

The NC is gaining ground, and can now be found not only in the spoken language, but also in student papers and in informal written registers of the language.

This example combines formal “Stylistic Fronting” of the past participle with the informal, nonstandard NC.

First nationwide study on the New Construction

Study 1: conducted 1999-2000, reported in Sigurjónsdóttir & Maling (2001) and

Maling & Sigurjónsdóttir (2002)   Questionnaire distributed to 1,731 tenth

graders (age 15-16) in 65 schools throughout Iceland, and 205 adult controls.

  The 1,731 tenth graders (age 15-16) constitute

45% of the children born in Iceland in 1984.

Sigurjónsdóttir & Maling (2001) survey question:

Is this sentence acceptable?

Það var beðið mig að vaska upp itEXPL was asked me-ACC to wash up

literally: ‘it was asked me to do the dishes’

intended: ‘I was asked to do the dishes’ (passive)

or ‘they asked me to do the dishes’ (active)

Is this sentence acceptable?

Það var beðið mig að vaska upp itEXPL was asked me-ACC to wash up

‘it was asked me to do the dishes’

N=200 N=1695

Study 2: conducted 2005-2007, and reported in Thráinsson et al. (to appear)

Study conducted on variation in a number of syntactic constructions in modern Icelandic

A subset of subjects throughout Iceland

(n=772) were tested on the New Construction. The subjects ranged from adolescents to seniors.

Second nationwide study on the New Construction

Is this sentence acceptable? Loks var fundið stelpuna eftir mikla leit. finally was found-neut girl.the-ACC after great search

‘The girl was finally found after a long search’ or ‘They finally found the girl after a long search’

Study 2: Thráinsson et al. (to appear) survey question:

In 2005-2007, Thráinsson et al. presented this example of the NC to 712 adolescents and adults in four age groups:

Study 2 (Thráinsson et al. to appear) ��Age-related variation in acceptance of New Construction ��

for this example; N=712

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

14-‐15 yrs. (N=200)

20-‐25 yrs. (N=179)

40-‐45 yrs. (N=168)

65-‐70 yrs. (N=165)

Impossible

Ques@onable

Natural

Summary: •  this change is widespread throughout Iceland; •  it is not due to contact (e.g. with Danish or English); •  it is not due to phonological or morphological changes. Therefore, it must be due to other system-internal factors.

Counterexample to Principle of Inertia

The New Construction seems to be a counter-example to the principle of “inertia” proposed by Keenan (1994, 2003) and developed by Longobardi (2001):

Syntax does not change on its own; syntactic change is triggered either by: (i) external forces (i.e., language contact), or (ii) some prior change in another domain, e.g. phonology or morphology.

Neither of these apply to the Icelandic NC


A marked option of Universal Grammar, namely a construction which assigns ACC without any (apparent) NOM, seems to have arisen spontaneously in the language without any prior relevant change in morphology or phonology.

We know that UG provides this marked option, cf. Polish & Ukrainian –no/to construction, Irish autonomous form, etc. (see M&S 2002, Maling 2006, inter alia)


The principle of inertia has considerable appeal because it reconciles the undoubted occurrence of syntactic change with the great accuracy with which children learn their native language syntax. It seems plausible that the output of acquisition could change only in reaction to a change in the information available to the first-language learner.


On-going research supports the hypothesis that the NC is acquired by young children and not adopted by adults. 46 of the subjects tested by M&S in 1999-2000 were retested in 2010; kids who acquire the NC do not seem to outgrow it (Thráinsson, Sigurjónsdóttir & Eythórsson, in progress).

Survey data on the acceptability of the NC show that the change is accepted by young people, but rejected by adults, strongly suggesting that the innovation is associated with language acquisition.


  Something must have happened c. 1950   What could it have been? Just an accident?   Or were there properties of Icelandic before

1950 that might have led children to reanalyze some existing construction?

We may no longer have access to the necessary information about the state of the language in the middle of the last century, when the NC first surfaced.

Some things we would like to know about the diachrony of the NC

2. Why did it arise in Icelandic, but not apparently in the other Germanic languages, especially Faroese and Norwegian, its closest relatives?

3. Is there anything about the preconditions for the change that supports one or another grammatical analysis of the NC?

1. Actuation problem: why did the NC surface in the mid-twentieth century and not earlier or later?

Possible models for the Icelandic New Construction

Það var hrint litlum strák itEXPL was pushed-neut.sg. little-DAT boy-DAT

‘A little boy was pushed’ (Eythórsson 2008, ex. 73b)

Það var dansað alla nóttina. itEXPL was danced all night ‘People danced all night’ (Maling & Sigurjónsdóttir 2002)

B. Impersonal Passive of an Intransitive Verb

Two possible sources for the emerging NC:

A. Impersonal or “Expletive Passive” of a Transitive Verb

Some passive clauses with a postverbal NP can only be analyzed as a Canonical Passive

a. Það var barinn strákur. itEXPL was hit-m.sg. boy-m.sg.NOM ‘A boy was hit’ b. Það var barin stelpa. itEXPL was hit-f.sg. girl-f.sg.NOM

‘A girl was hit’

Note: masculine nouns and feminine singular nouns have distinct NOM/ACC forms.

Syntactic ambiguity of Expletive Passives

a. Það var skammað lítið barn. itEXPL was scolded-neut. little child-neut.sg ‘A little child was scolded’ (Eythórsson 2008, ex.73a) b. Það var hrint litlum strák. itEXPL was pushed-neut little-DAT boy-m.sg.DAT

‘A little boy was pushed’ (Eythórsson 2008, ex.73b)

But many sentences in standard Icelandic can be analyzed as either the Canonical Passive or the NC (Sigurjónsdóttir & Maling 2001:128; Thráinsson 2007:276; Eythórsson 2008)

Note: neuter nouns have the same form for NOM/ACC in both singular and plural; feminine nouns are nondistinct in the plural.


What are the implications of this fact for language acquisition?

  For such clauses, the child cannot tell that the postverbal NP is NOM

  As a result, she may hypothesize that the postverbal NP is actually an object, not a postposed subject.

We have a case of ‘misanalysis’ or ‘reanalysis.’ Default rule: a postverbal NP gets marked ACC

Fact: neuter nouns have the same form for NOM/ACC in both singular and plural; feminine forms are nondistinct in the plural.

Implications for language acquisition

“...there is considerable overlap between the two dialects. That means, of course, that in the primary linguistic data (PLD) available to a child acquiring the language there is a lot of ambiguity even if the data all come from speakers of the standard dialect. That is a typical situation for ‘misanalysis’ by children, an important source of language change” (Thráinsson 2007:276)

Implications for language acquisition

a. “Once the postverbal NP has been reanalyzed as an object that is assigned structural accusative case, the New Passive emerges.” (Eythórsson (2008:212-213)

b. “Since the NP is not a subject but an object, the DE no longer applies.”

One hypothesis: the New Construction is an impersonal “expletive passive” that has lost the definiteness constraint that exists in the standard language (Eythórsson 2008)

The direct object is marked ACC because it is a “non-promotional” passive; as a passive, it has an empty [e] subject.


Frequency data from the IcePaHC corpus supports the suggestion that morphological nondistinctness is a source of syntactic misanalysis.

Neuter nouns have the same form for NOM/ACC in both singular and plural; feminine forms are nondistinct in the plural. Only feminine sg. and masculine nouns have distinct NOM/ACC forms.

(data provided by Joel Wallenberg (p.c.))

The Icelandic Parsed Historical Corpus (IcePaHC)

Wallenberg, Joel C., Anton Karl Ingason, Einar Freyr Sigurðsson and Eiríkur Rögnvaldsson. 2010. Icelandic Parsed Historical Corpus (IcePaHC). Version 0.2. http://www.linguist.is/icelandic_treebank

  Goal: 1 million words by August, 2011   currently c. 341,000 words   Time period covered: c. 1100-1900 Supported by:   Icelandic Research Fund (RANNÍS) (#090662011)   U.S. National Science Foundation (NSF) (#OISE-0853114)   University of Iceland Research Fund (Rannsóknasjóður HÍ)

Wallenberg, Joel C., Anton Karl Ingason, Einar Freyr Sigurðsson and Eiríkur Rögnvaldsson. 2010. Icelandic Parsed Historical Corpus (IcePaHC). Version 0.2. http://www.linguist.is/icelandic_treebank

Ambiguity of postverbal NP in passive

A search for BE > passive-participle > subject, regardless of what occupied the first position in the clause, using CorpusSearch, found 24 subordinate passive clauses with a post-participle NP. (Joel Wallenberg, p.c., Jan. 2011)

Results: •  7/24 are clearly NOM; two of them are definite •  11/24 are ambiguous between NOM/ACC •  6/24 are Dative; all six are definite

•  only about ¼ of the postverbal NPs in passive clauses were unambiguously NOM •  roughly half were ambiguous between NOM/ACC

b. þegar frá eru teknir biskupsstólarnir when from are taken-m.pl. bishop.seats-m.pl-NOM ‘when the bishoprics are excluded’

(Sturlunga Saga, mid-13th century)

a. úr er tekinn raddarstafur out is taken letter-m.sg.NOM

‘a letter is taken out’ (c.1100, First Grammatical Treatise)

Assumption: in an expletive passive clause, an NP in post-participle position is unambiguously a subject only if it is marked NOM (we exclude ditransitives from discussion)


2 examples of unambiguous nominative NPs:

og er þó minnkað atkvæði nafns þeirra and is though reduced-n.sg syllable-n.sg. name-G their-G ‘and one syllable of their name is however reduced’ (c. 1100, First Grammatical Treatise)



Search for BE > passive-participle > subject, regardless of what occupied the first position in the clause, using CorpusSearch, found 24 subordinate passive clauses with a post-participle NP

many expletive passives open to ‘misanalysis’

upp var lokið myrkvastofunni up was locked-3sg.neut prison.the-DAT

(1525, Erasmus Saga)



Search for BE > passive-participle > subject, regardless of what occupied the first position in the clause, using CorpusSearch, found 24 subordinate passive clauses with a post-participle NP.


BUT, we do not yet see the New Construction; there is only one unambiguous example of an ACC NP with a passive participle in the IcePaHC corpus.

•  only about ¼ of the postverbal NPs in passive clauses were unambiguously NOM •  roughly half were ambiguous between NOM/ACC

If this data is representative of the data available to the language learner, then the expletive passive construction is a possible source of misanalysis.

En þá kvöl, sem eg hafði að bera af kitlum, sem eg hafði í yljum og tám, verður ei af mér útmálað. “But the torment which I had to endure from the tickling that I had in the soles of my feet and toes will not be described by me”

Earliest attested example of NC? En þá kvöl, …verður ei af mér útmálað but the-ACC torment-fem,…becomes not by me out.painted-neut.sg “The torment …will not be described by me.”

(1791, Jón Steingrímsson, diary/autobiography)

Plausible story for reanalysis – but…

“The fact remains that there do not seem to be any unambiguous examples of the New Passive containing either full NPs in the accusative or pronouns until the mid-20th century (Maling & Sigurjónsdóttir 2002:129)”

Problem #1: Actuation Problem: why now? this morphosyntactic ambiguity has been true throughout the history of Icelandic, yet...

(Eythórsson 2008:212)

Eythórsson claims that “the Norwegian det-passive is a close parallel to the New Passive in Icelandic in that the postverbal argument is an object rather than a subject ... the difference is that the direct object NP must generally be indefinite in Norwegian” (Eythórsson 2008:206)

The claimed parallelism is not convincing: The argument that the postverbal NP receives ACC case in Norwegian is theory-internal. In the Norwegian det-passive, the postverbal NP is a postposed indefinite subject. Personal pronouns are excluded in Norwegian, whereas they are common in the Icelandic NC.

Problem #2: why in Icelandic but not Faroese or Norwegian?


Because the Indirect Object of a ditransitive verb can be definite, Eythórsson (2008) claims that Norwegian ditransitives show the predicted lack of the DE ; note however, that only the Indirect Object can be definite, the direct object must be indefinite.

Det vart overrekt vinnaren ein pokal /*pokalen. itEXPL was given the.winner a cup/ *the.cup ‘The winner was given a cup/*the cup’ ( = Eythórsson 2008, ex. (66b))

Det var lagt eit document/*det framfor oss. it was placed a document/ *it before us ‘A document was placed before us’ (=Eythórsson 2008, ex. 66a)



The det-passive construction is an existential with a (personal) passive base. The construction targets the Direct Object, the theme, and the definiteness Effect follows. The Indirect Object is inert.

CONCLUSION: it is misleading to describe the Icelandic NC as a “parallel development” to the Norwegian det-construction. We know of no evidence that Norwegian is extending it along the same lines to transitive verbs, apart from the simple reflexive seg.

Another possible model for the Icelandic New Construction

The surveys included examples of the New Construction and Canonical Passives, but they also included examples of the traditional “Impersonal Passive,” a construction which all linguists (including us) considered to be passive. (Sigurjónsdóttir & Maling 2001; Maling & Sigurjónsdóttir 2002, etc.)

Það var dansað alla nóttina. itEXPL was danced all night ‘People danced all night’

Another possible model: Impersonal Passive of Intransitive Verbs

Our survey (M&S 2002) included 200 adult controls. The most surprising result of our survey was that for many of the adults, the traditional Impersonal Passive displayed two of the syntactic properties that we had identified as being associated with being Active as opposed to Passive: (i) control of subject oriented participles, and (ii) bound anaphors

Það var dansað alla nóttina. itEXPL was danced all night ‘People danced all night’

Another possible model: Impersonal Passive of Intransitive Verbs

Control of subject-oriented adjuncts is a property of actives. Do we find this syntactic property in the traditional Impersonal Passive?

Það var komið skellihlæjandi í tímann itEXPL was come laughing into class

‘People came into class laughing’ (=Maling & Sigurjónsdóttir 2002, ex. (37a))

Note: these participial adjuncts are verbal and not simple adverbs; they can take complements.

Canonical Passive – implicit agent cannot control these adjuncts b. *Börnin voru vakin kl. 7, raulandi lítinn lagstúf. the.children were awakened at 7, humming little song.bit

Active – subject-oriented participial adjunct a. Ég vakti börnin kl. 7, raulandi lítinn lagstúf I woke the.children at 7a.m., humming little song.bit “I woke the children at 7a.m., humming a little piece of a song”

Control of subject-oriented adjuncts is a property of actives.

These participial adjuncts can take complements:

New Construction – allows subject-oriented participial adjuncts c. Það var vakið börnin kl. 7, raulandi lítinn lagstúf. itEXPL was awakened the.children at 7, humming little song.bit

0%

10%

20%

30%

40%

50%

60%

70%

80%

Beyond Reykjavík

Inner Reykjavík

Adults

Adjunct OK Imp. Passive

% Accept New Const.

Acceptance of subject-oriented adjuncts in�� Impersonal Passives vs. overall rates of acceptance for��

the New Construction (M&S 2002)

The Impersonal Passive is showing the same active syntactic property as the New Construction, but at higher levels of acceptance! So it is leading the innovation, not following.��(See also Barðdal & Molnár 2003; Sigurðsson & Egerland 2009)

B. Control of bound anaphors is a property of actives ��The nationwide surveys showed that for many adults, reflexive verbs can occur in Impersonal passives.

Svo var bara drifið sig á ball. then was just hurried REFL-ACC to dance ‘Then everyone just hurried off to the dance’ (Maling &Sigurjónsdóttir 2002, ex. 30a)

This is not the New Construction, it is the traditional Impersonal Passive with a reflexive verb. Reflexive verbs are known to pattern with intransitive verbs in many languages. Many Icelandic speakers find it acceptable to include a reflexive or other bound anaphor.

See also Sigurðsson (1989:355, fn. 60); Barðdal & Molnár (2003)

Svo var bara drifið sig á ball. then was just hurried REFL-ACC to dance ‘Then everyone just hurried off to the dance’ (M&S 2002, ex 30a)

This sentence was included in both surveys; data below from Thráinsson et al. show the age-related variation clearly.

Compare this level of acceptability with the levels of acceptability by age that we saw for the New Construction:

0%

10%

20%

30%

40%

50%

60%

70%

80%

14-‐15 20-‐25 40-‐45 65-‐70

Reflexive OK Imp.Passive

% Accept New Const.

Again, the Impersonal Passive is showing the same active syntactic property as the New Construction, but at higher levels of acceptance!

Impersonal passives allow control of adjuncts and control of �� reflexives at greater rates for younger people, but the shift�� probably started within the last century.

“I have not been able to find any cases of [Impersonal passives] with reflexive verbs in Old Icelandic; ….Thus, the reflexive [Impersonal passive] seems to be an innovation of Modern Icelandic which is increasingly gaining ground and is accepted by many speakers who do not accept the [New Construction] … (cf. Maling and Sigurjónsdóttir, p.122).”


Acceptance of Bound Anaphora in the Impersonal Passive is a 20th Century Phenomenon.

Based on a corpus search on an open-access digital library (timarit.is) which hosts digital editions of newspapers and magazines from the 17th century to the early 21st century, Einar Freyr Sigurðsson observed that “you only get sporadic examples of reflexive impersonal passives in the earlier periods, but after about 1960 or so … the number of examples increases significantly.”

(p.c., Einar Freyr Sigurðsson, August, 2010)

Acceptance of Bound Anaphora in the Impersonal Passive is a 20th Century Phenomenon.

Characteristic ‘S’-shaped curve

See Kroch (1989) “Reflexes of Grammar in Patterns of Language Change,” Language Variation and Change 1:199-244.

“A given change begins quite gradually; after reaching a certain point (say, twenty per cent), it picks up momentum and proceeds at a much faster rate; and finally tails off slowly before reaching completion. The result is an ∫-curve…” C.-J. Bailey, Variation and Linguistic Theory, (1973:77)

Passives of reflexive verbs have been considered marginal by some researchers, e.g. Halldór Ármann Sigurðsson (1989:355, fn. 60); Eythórsson asserts that “the ImpC with reflexives (‘reflexive passive’) is rather marginal and is hardly robust enough to be a model for the NC transitives” (Eythórsson (2008:215, emphasis added).

But elsewhere in that same paper, Eythórsson says that he himself and most speakers he consulted find them acceptable. See also Barðal & Molnár 2003. Linguists who are now in their fifties tell us that they have noticed an increase in the occurrence of reflexive impersonal passives over the last two decades or so, and believe that their acceptability judgments may have changed (Eiríkur Rögnvaldsson, p.c.).

Impersonal Passive of Reflexive verbs is a 20th C. phenomenon

About half of the adult population is accepting active syntactic properties as part of the Impersonal Passive. What could explain this?

This is a syntactic change that had NOT been noticed.

Native speakers of Icelandic have all noticed the eruption of the New Construction, but they had not noticed the slowly shifting tectonic plates that have led up to the New Construction.

What is the nature of this much more subtle change?

Hypothesis: Impersonal passives of intransitive verbs are in principle syntactically ambiguous between active and passive. ��(Sigurjónsdóttir & Maling 2001; Maling & Sigurjónsdóttir 2002)

Impersonal Passive: [e] [VP var dansað] Grammar 1

Impersonal Active: [proarb] [VP var dansað] Grammar 2

i.e. the data underdetermines the analyses

Hypothesis: Impersonal passives of intransitive verbs are in principle syntactically ambiguous between active and passive. �� (Maling & Sigurjónsdóttir 2002)

Hypothesis: roughly half of adult speakers responding to our survey analyzed the traditional Impersonal Passive as a passive construction. The other half analyzed it as active, i.e. having a syntactically accessible null subject. This makes bound anaphors and adjuncts possible.

So if a speaker has the Impersonal Passive represented as

[e] [VP var dansað] then she will not accept control of adjuncts because there is nothing to control them.

Some anedotal evidence supports this claim:

��When asked to consider this example:��


‘People came into class laughing’ ��

�� one speaker in her 70s said: “Það vantar einhvern.”

Someone is missing. �� [e] [VP var komið í timann]

Hypothesis: Impersonal passives of intransitive verbs are in principle syntactically ambiguous between active and passive. ��(Sigurjónsdóttir & Maling 2001; Maling & Sigurjónsdóttir 2002)

This makes an interesting prediction: those adults who accept the adjuncts in the Impersonal Passive should be more likely to accept the bound anaphors in the Impersonal Passive, because their syntactic representation of the construction provides for both.

��

[proarb] [VP var dansað]

This prediction is borne out: For both adolescents and adults, acceptance of subject-oriented participles is significantly correlated with acceptance of reflexives in impersonal passives. Adolescents: r = 0.43 p< .001 (n=1693) Adults: r = 0.53 p< .001 (n=199) Hypothesis: speakers who accept subject-oriented participles and reflexives have an active representation of the “Impersonal Passive,” while speakers who reject subject-oriented participles and reflexives have a passive representation of the “Impersonal Passive.”

Haspelmath: Syntactic ambiguity of ��“impersonal passives” of intransitive verbs

“Notionally, generalized subject constructions are close to the passive, especially in that the agent is backgrounded.” (p. 49)

Haspelmath, Martin: 1990, “The grammaticization of passive morphology,” Studies in Language 14.1, 25-72

“The difference between passive and desubjective [active clause with no subject] is of a syntactic rather than a semantic nature.” (p. 58)

“...intransitive desubjectives are indistinguishable from passives of intransitive verbs, so transitive desubjectives are the crucial case.” (Haspelmath 1990:35)

Possible stages of change in the grammatical system:

Stage 1. Impersonal passives occur only with intransitive verbs (e.g. dansa ‘dance’) (Icelandic before c. 1900) (Grammar 1)

Stage 2. Impersonal passives start to occur with reflexive verbs in the 20th century, as some speakers reanalyze the Impersonal Passive as a syntactically active construction with a proarb subject (Grammar 2).

Stage 3. For Grammar 2 speakers, Impersonal “passives” occur with all transitive verbs, with ACC on retained object. This is what we recognize as the�� “New Construction”

1.  Within-subject analysis of survey data to determine grammars of individual speakers

2. Longitudinal studies (in progress)

3. Sociolinguistic studies

4. Discourse context of NI vs. canonical passive

5. Comparative studies of other Germanic languages

6. Corpus studies: annotated corpora under construction

Where do we go from here?

Takk fyrir!��Thank you for listening

Acknowledgments: U.S. National Science Foundation, SBR-9223725 (pilot study in 1997) Lýðveldissjóður, Republic of Iceland Vísindasjóður RANNÍS (Icelandic Science Foundation) Rannsóknasjóður Háskólans (University of Iceland Research Fund)

Apparent German parallel? See M&S 2002:131, fn.20

Syntactic Properties of the Norwegian Impersonal Passive – the selv ‘self’ anaphor in (b) is not allowed

a. Det ble danset av alle og enhver i bygda.

it was danced by one and all in the.village

b.  Det ble låst seg (*selv) inn i fabrikken.

it was locked REFL (self) inside in the.factory

c. *Det ble danset leende/gråtende/full/fulle/fullt.

it was danced laughing/crying/drunk

d. *Under krigen ble det forsvunnet ofte uten spor.

in the.war was it disappeared often without a.trace

Another possible model in Icelandic for the reanalysis

Það var verið [að borða fisk] itEXPL was been to eat fish-ACC

‘People were eating fish’ Thráinsson (2007:429)

Icelandic has an unusual combination of auxiliary be + past participle which is not passive but active in meaning:

verið – passsive participle? perfect participle? supine?

These impersonal passives of aspectual vera ‘to be’ are part of the standard language, and may be a possible model for the NC (M&S 2002:134; Maling 2006:218)

Impersonal passives of aspectual vera ‘to be’ are part of the standard language, and may be possible model for the NI

a. “Í gær þegar það var gefið mér lýsi , þá…” yesterday when it was given me lýsi, then... (girl, age 4;4) Standard language: b. Í gær þegar var verið að gefa mér lýsi, þá… yesterday when was been to give me lýsi, then... ‘Yesterday when they were giving me cod liver oil,

then....’

  „In passing, it is worth pointing out that the impersonal passive is extremely common in Icelandic (cf. Friðjónsson 1987), for example much more common than in German, I believe (German very typically replacing it by an active man- ‚one‘ construction).“ (Sigurðsson 1989:162)

  Note: Sigurðsson includes verbs with PP-complements in this category: Stundum var [e] hlegið að rádherranum. (= Sigurðsson 1989:162, ex.2a)

sometimes was laughed at the.minister   Barðdal & Molnár (2003:243f) categorize them separately as a

“prepositional passive” Það var hlegið að skemmtikraftinum. (=Barðdal & Molnár 2003:244, ex. 26b)

itEXPL was laughed at the.comedian

Impersonal passives are relatively frequent

Agentive by-phrases are infrequent in Icelandic

by-phrases are infrequent in Icelandic data from Joel

Insert quote from Tolli or JGJ??

“...the thematic restrictions on Icelandic passive are more strict than those of the English passive” (Thráinsson 2007:257)

“In the active voice, the subject of a transitive verb like eyðileggja ‘destroy’ can be a natural force such as fire, storm, flood or avalanche. However, in a personal passive..., the understood agent must be a human.” (Maling & Sigurjónsdóttir 2002:132)

Joel Wallenberg compared the frequency of passive vs. active for transitive verbs in the translations of the New Testament Book of John by Oddur Gottskálksson, Luther, and Tyndale.

Language: Ice Ger Eng Voice: passive 84 97 190 active 1100 1071 976 frequency 0.071 0.083 0.163 The difference between Icelandic and German is not statistically significant, but the difference between English and the others is.

(Joel Wallenberg, p.c. 4 February, 2011) Note: passives of transitive or ditransitive verbs; excludes impersonal passives of intransitive

verbs

Passives are relatively infrequent in Icelandic

Wallenberg compared the frequency of passive vs. active for the Icelandic and English translations of Acts 1-17, corresponding samples.

Lang Ice Eng Voice passive 95 145 active 714 570 Freq. 0.117 0.203 The same effect holds: the frequency of passive voice is roughly

twice as high in English as in Icelandic Note: passives of transitive or ditransitive verbs; excludes impersonal passives of intransitive

verbs


Wallenberg then checked to see if this effect held between the full corpora of Early Modern English and the IcePaHC.

PPCEME and IcePaHC: Language Ice Eng Voice passive 1621 17273 active 9851 40096 Freq. 0.141 0.301 The same effect holds: the frequency of passive voice is roughly

twice as high in English as in Icelandic Note: count includes passives of transitive or ditransitive verbs; excludes impersonal passives

of intransitive verbs


Syntactic Property Polish (active)

Ukrain. (passive)

Agentive by-phrase possible

* Yes

Control of subject-oriented adjuncts

Yes *

Non-agentive “Unaccusative” verbs

are possible

Yes

*

Bound anaphors in underlying object

position

Yes

*

Syntactic behavior of –no/to construction in Polish vs. Ukrainian (Maling & Sigurjónsdóttir 2002)

The New Impersonal is parallel to development of the ��–no/to-construction in Polish, an active impersonal with a thematic proarb subject

(Maling &Sigurjónsdóttir 1997, 2002; Maling 2006)

The New Impersonal is “comparable to the –no/to construction in Ukrainian, a passive preserving structural accusative case”


Impersonal Passive


‘They came into class laughing’

Það var lesið minningargreinina grátandi itEXPL was read the.memorial.article crying

‘They read the memorial article crying’ ‘The memorial article was read, crying’

New Construction

0% 10% 20% 30% 40% 50% 60% 70% 80%

Adolesc. beyond

Reykjavík

Adolesc. Inner

Reykjavík

Adults

How do these rates compare with overall acceptance of ��the New Construction?

% Acceptance for control of adjuncts in NC, compared with acceptance rates for the NC overall

% Acceptance for control of adjuncts in NC, compared with acceptance rates for the NC overall

So control of subject-oriented adjuncts is a property that is acceptable to most people who accept the New Construction.

The New Construction with a possessive reflexive (M&S 2002)

Á kvöldin var skoðað tölvupóstinn sinn. in evenings was viewed e-mail-ACC refl ‘In the evenings people checked their own e-mail’ ‘In the evenings their own e-mail was checked.’

Notice in the analogous passive sentence in English, the bound reading is blocked: In the evenings their e-mail was checked.

So if acceptance of bound anaphora in the NC is evidence that the NC is active, we would predict that speakers will NOT allow bound anaphora in CANONICAL Passives.

Tölvupósturinn (*sinn) var skoðaður á kvöldin. e-mail-m.sg.NOM (*REFL) was checked-m.sg. in the.evening

‘(*their own) e-mail was checked in the evenings’

Due diligence: Can the implicit Agent in a Canonical Passive bind a reflexive?

No: Canonical passives are bad with reflexives, as expected.

Maling, Joan & Sigríður Sigurjónsdóttir (2002) “The New

Impersonal Construction in Icelandic,” Journal of Comparative Germanic Linguistics 5:97-142.

Maling, Joan (2006) “From passive to active: syntactic change in progress in Icelandic,” in Lyngfelt & Solstad (eds.) Demoting the Agent: Passive, middle and other voice phenomena, John Benjamins, pp. 197-223.

  research supported by grants fromVísindasjóður

Rannsóknarráðs Íslands (RANNÍS), Rannsóknarsjóður Háskóla Íslands, Lýðveldisjóður and in part by NSF grant BCS-9223725 to Brandeis University.

Acceptability rates for the sentence Svo var bara drifið sig einn á ball. so was just gone-neut REFL alone-m.sg. to dance ‘So people just went to the dance alone’ Elsewhere Inner Rvík Adults 60% 48% 22.5% (previously unreported results from the Sigurjónsdóttir

& Maling 1999-2000 study; the acceptance rate for adolescents in Elsewhere ranged from 52-70%)

Reflexive impersonal passive with adjunct einn ‘alone’

Nothing personal? A system-internal syntactic change in ...kroch/hist-pdf/maling-goett11.pdf · Sigurjónsdóttir & Maling (2001) and Maling & Sigurjónsdóttir (2002) Questionnaire

Documents